Minor HW corrections

hw_sentiment.ipynb

@@ -849,8 +849,8 @@
    "outputs": [],
    "source": [
     "def predict_one_rnn(text):\n",
-    "    # Singleton list of feature dicts:\n",
-    "    feats = [rnn_experiment['phi'](text)]\n",
+    "    # List of tokenized examples:\n",
+    "    X = [rnn_experiment['phi'](text)]\n",
     "    # Standard `predict` step on a list of lists of str:\n",
     "    preds = rnn_experiment['model'].predict(X)\n",
     "    # Be sure to return the only member of the predictions,\n",

hw_wordrelatedness.ipynb

@@ -49,7 +49,7 @@
     "\n",
     "Word similarity and relatedness datasets have long been used to evaluate distributed representations. This notebook provides code for conducting such analyses with a new word relatedness datasets. It consists of word pairs, each with an associated human-annotated relatedness score. \n",
     "\n",
-    "The evaluation metric for each dataset is the [Spearman correlation coefficient $\\rho$](https://en.wikipedia.org/wiki/Spearman%27s_rank_correlation_coefficient) between the annotated scores and your distances, as is standard in the literature. Since the train and test datasets contain both similarity and relatedness datasets, we will also report separate scores for each of these sub-tasks as well as an overall score.\n",
+    "The evaluation metric for each dataset is the [Spearman correlation coefficient $\\rho$](https://en.wikipedia.org/wiki/Spearman%27s_rank_correlation_coefficient) between the annotated scores and your distances, as is standard in the literature.\n",
     "\n",
     "This homework ([questions at the bottom of this notebook](#Homework-questions)) asks you to write code that uses the count matrices in `data/vsmdata` to create and evaluate some baseline models. The final question asks you to create your own original system for this task, using any data you wish. This accounts for 9 of the 10 points for this assignment.\n",
     "\n",